Monday, 30 June 2025
30.6 C
Singapore
34.4 C
Thailand
21.6 C
Indonesia
29.7 C
Philippines

DeepSeek’s R1 model was found to be highly vulnerable to jailbreaking

DeepSeek’s R1 AI model is reportedly more vulnerable to jailbreaking than other AI systems, raising concerns about its ability to produce harmful content.

The latest artificial intelligence model from DeepSeek, the Chinese AI company making waves in Silicon Valley and Wall Street, is more susceptible to manipulation than other AI models. Reports indicate that DeepSeek’s R1 can be tricked into generating harmful content, including plans for a bioweapon attack and strategies to encourage self-harm among teenagers.

Security concerns raised by experts

According to The Wall Street Journal, DeepSeek’s R1 model lacks the robust safeguards seen in other AI models. Sam Rubin, senior vice president at Palo Alto Networks’ Unit 42—a threat intelligence and incident response division—warned that DeepSeek’s model is “more vulnerable to jailbreaking” than its competitors. Jailbreaking bypasses security filters to make an AI system generate harmful, misleading, or illicit content.

The Journal conducted its tests on DeepSeek’s R1. It was able to manipulate it into designing a social media campaign that, in the chatbot’s own words, “preys on teens’ desire for belonging, weaponizing emotional vulnerability through algorithmic amplification.”

AI model produces dangerous content

Further testing revealed even more concerning results. The chatbot reportedly provided instructions for executing a bioweapon attack, drafted a pro-Hitler manifesto, and composed a phishing email embedded with malware. In comparison, when the same prompts were tested on ChatGPT, the AI refused to comply, highlighting the significant security gap in DeepSeek’s system.

Concerns about DeepSeek’s AI models are not new. Reports suggest that the DeepSeek app actively avoids discussing politically sensitive topics such as the Tiananmen Square massacre or Taiwan’s sovereignty. Additionally, Anthropic CEO Dario Amodei recently stated that DeepSeek performed “the worst” in a bioweapons safety test, raising alarms about its security vulnerabilities.

Hot this week

Android 16 to alert you if your phone connects to a fake cell tower

Android 16 will warn you if your phone connects to a fake tower, helping protect your calls, texts, and location from silent spying.

Meta may buy PlayAI to boost its voice cloning technology

Meta may buy AI voice cloning startup PlayAI to expand lifelike voice features in its apps, smart glasses, and AI assistants.

ASUS V500 Mini Tower (V500MV): Efficient desktop power for home and office

ASUS V500 Mini Tower (V500MV) is a quiet, space-efficient desktop with Intel Core i7, dual storage, and tool-free access, built for daily productivity.

AWS opens innovation hub in Singapore to drive cloud and AI adoption across Asia Pacific

AWS opens first Innovation Hub in Asia Pacific to accelerate digital transformation with cloud and AI for regional businesses.

HDMI 2.2 launches with support for 16K video and 96Gbps cables

HDMI 2.2 supports 16K video, 96Gbps cables, and audio sync upgrades, setting a new standard for future-ready home entertainment systems.

Cheapest SIM-only plans in Singapore 2025: Flexible, contract-free mobile data

Compare the cheapest SIM-only plans in Singapore for 2025, with up to 1TB data, 5G access, roaming, and no-contract options from S$8/month.

Android 16 to alert you if your phone connects to a fake cell tower

Android 16 will warn you if your phone connects to a fake tower, helping protect your calls, texts, and location from silent spying.

Runway moves into gaming with new AI platform Game Worlds

Runway launches Game Worlds, an AI platform aiming to reshape game creation and expand its success from film into the gaming industry.

TikTok trials new ‘bulletin boards’ to rival Instagram’s broadcast channels

TikTok is testing bulletin boards, a new feature similar to Instagram's broadcast channels, for direct creator-to-fan updates.

Related Articles

Popular Categories